Spectra and physical properties of Taurid meteoroids

Pavol Matlovica, Juraj Totha, Regina Rudawskab, Leonard Kornosa

aFaculty of Mathematics, Physics and Informatics, Comenius University, Bratislava,Slovakia

bESA European Space Research and Technology Centre, Noordwijk, Netherlands

Abstract

Taurids are an extensive stream of particles produced by comet 2P/Encke,which can be observed mainly in October and November as a series of me-teor showers rich in bright fireballs. Several near-Earth asteroids have alsobeen linked with the meteoroid complex, and recently the orbits of two car-bonaceous meteorites were traced to the center of the stream, raising in-teresting questions about the origin of the complex and the composition of2P/Encke. Our aim is to investigate the nature and diversity of Taurid mete-oroids by studying their spectral, orbital, and physical properties determinedfrom video meteor observations. Here we analyze 33 Taurid meteor spectracaptured during the predicted outburst in November 2015 by stations in Slo-vakia and Chile, including 14 multi-station observations for which the orbitalelements, material strength parameters, dynamic pressures, and mineralog-ical densities were determined. It was found that while orbits of the 2015Taurids show similarities with several associated asteroids, the obtained spec-tral and physical characteristics point towards cometary origin with highlyheterogeneous content. Observed spectra exhibited large dispersion of ironcontent and significant Na intensity in all cases. The determined materialstrengths are typically cometary in the KB classification, while PE criterionis on average close to values characteristic for carbonaceous bodies. Thestudied meteoroids were found to break up under low dynamic pressures of0.02 - 0.10 MPa, and were characterized by low mineralogical densities of 1.3- 2.5 g cm-3. The widest spectral classification of Taurid meteors to date ispresented.

Keywords: meteor shower, meteoroid stream, comet, asteroid, Taurids,

Email address: matlovic@fmph.uniba.sk (Pavol Matlovic)

Preprint submitted to Planetary and Space Science December 30, 2016

2P/Encke

1. Introduction

The Taurid complex is an extensive population of bodies associated withcomet 2P/Encke and certainly one of the most interesting structures in theSolar System. It is the widest known meteoroid stream in the Solar System,and has been linked with several catastrophic incidents including the Palae-olithic extinctions (Napier, 2010) and the Tunguska event of 1908 (Kresak,1978). The low inclination of the orbit of the stream causes gravitationalperturbations of inner solar system planets (Levison et al., 2006), resultingin the diffuse structure of the Taurid complex. The Earth encounters differ-ent parts of the stream annually from September to December, giving riseto several meteor showers, most significantly the Northern and the South-ern Taurids peaking in late October and early November. Owing to the widespread of the Taurid stream, the zenithal hourly rates of the resulting meteorshowers usually do not exceed 5. However, the observed activity of the showeris known to be rich in bright fireballs, which induced discussions about thepotential of Taurids in producing meteorites (Brown et al., 2013; Madiedo etal., 2014; Brown et al., 2016). While the orbits of most meteoroids are quitedispersed, it is still likely that the Taurid stream has a narrow and dense coreconsisting of particles concentrated near the orbit of the stream’s parent ob-ject. It is often inferred that strong bombardment episodes have resulted atepochs when the material of the stream’s core reached Earth intersection.

Although comet 2P/Encke is generally considered as the parent object ofthe Taurids (Whipple, 1940), various small near-Earth objects (NEOs) andrecently even two instrumentally observed carbonaceous meteorite falls havebeen linked with the orbit of the stream (Haack et al., 2011; Haack et al.,2012). 2P/Encke is a short-period comet moving on an orbit dynamicallydecoupled from Jupiter with an orbital period of 3.3 years. The peculiarorbit of the 2P/Encke raises interesting questions relating its origin and theorigin of the Taurid complex. Today, the most supported hypothesis claimsthat comet 2P/Encke and several other bodies including large number ofmeteoroids were formed by a fragmentation of an earlier giant comet. Thetheory was first suggested by Whipple (1940), and later elaborated by Asheret al. (1993) who argued that the Taurid meteoroid streams, 2P/Encke, andthe associated Apollo asteroids were all formed by major comet fragmentation

2

20 to 30 ky ago. Napier (2010) suggested that the debris of this fragmentationevent could have caused the Palaeolithic extinctions followed by the returnto ice age conditions.

The broad structure of the Taurid complex was studied by several au-thors with over 100 of NEO candidates associated with the Taurid complex.Porubcan et al. (2006) identified 15 different sub-streams of the complexand by applying stricter criteria for generic relations found associations with9 NEOs. The presence of these bodies in the Taurid complex implies thepossibility that part of the Taurid meteoroid population might be producedas the decay or impact products of the associated asteroids. The connectionto carbonaceous chondrites represented by Maribo and Sutter’s Mill mete-orites also needs to be further investigated. We expect significantly differentspectral and structural properties of such bodies in comparison to fragilecometary meteoroids originating in 2P/Encke. So far, the efforts in find-ing traces of the common origin of the largest Taurid complex bodies gaverather sceptical results. Popescu et al. (2014) studied the spectral proper-ties of the largest asteroids associated with the Taurid complex, but foundno evidence supporting mutual relation. Similarly, the spectroscopic andphotometric measurements of Tubiana et al. (2015) found no apparent linkbetween comet 2P/Encke, the Taurid complex NEOs, and the Maribo andSutter’s Mill meteorites.

The Taurid meteor shower occasionally exhibits enhanced activity dueto a swarm of meteoroids being ejected by the 7:2 resonance with Jupiter(Asher, 1991). The outburst of meteors observed in 2015 was anticipatedby the model of Asher & Izumi (1998), which previously predicted enhancedactivities of Taurids in 1998, 2005 and 2008. Some features of the Taurid me-teor shower activity suggest that the Taurid swarm exist only in the southernbranch (Southern Taurids) and not in the northern branch (Shiba, 2016).

The two main branches of the Taurid complex, the Northern and theSouthern Taurids are well observed meteor showers with established orbitalcharacteristics, which clearly trace them to comet 2P/Encke. The majority ofTaurid shower analyses have been focused on orbital properties of the stream,with only several Taurid meteor spectra observed before the outburst of 2015(e.g. Srirama Rao & Ramesh, 1965; Madiedo et al., 2014). Borovicka et al.(2005) analyzed six Taurid emission spectra observed by low-resolution videospectrograph, identifying three Na-enhanced meteoroids, while another threeTaurids have been classified as normal type. Using the same instrumentation,Vojacek et al. (2015) observed three more normal type Taurids including one

3

meteoroid with lower iron content. The determined spectral characteristicsof normal to enhanced sodium content and lower iron content would suggestcometary parent body. Asteroidal meteoroids are expected to be depleted involatile sodium by the processes of space weathering (e.g. Borovicka et al.,2005; Trigo-Rodrıguez & Llorca, 2007). Determined spectral properties giveus valuable input into the studies of the origin of meteoroids; however, fordoubtless differentiation between cometary and asteroidal particles, preciseorbital and ideally physical characteristics must be obtained.

Besides the two major Taurid meteor showers, there are several smallermeteor streams associated with the Taurid complex. Most notably, this in-cludes the Piscids, Arietids, chi Orionids, and the daytime showers of betaTaurids and zeta Perseids, encountered by the Earth in June and July. Thebeta Taurids and zeta Perseids have been shown to be the daytime twinbranches to the Southern and the Northern Taurids respectively. The possi-bility of Taurid sub-streams being produced by a different parent object wasdiscussed by Babadzhanov (2001), who found shower associations to each ofthe Taurid complex asteroids and interpreted them as the evidence for thecometary origin of these asteroids. Recently, Olech et al. (2016) found veryclose resemblance between the orbits of two 2015 Southern Taurid fireballsand the orbits of 2005 UR and 2005 TF50 asteroids. Precise observations ofTaurid fireballs observed during the 2015 outburst were examined by Spurnyet al. (2016), who emphasized the orbital resemblance of the 2015 Taurids tothe orbits of asteroids 2015 TX24 and 2005 UR. They argue that the outburstmay have been caused by Taurid filaments associated with these asteroids.

All of the previous associations between different sub-streams of the Tau-rids and the Taurid complex asteroids were based on the orbital similarities.Certainly, studying the spectral and physical properties of the Taurid me-teoroids could extend our understanding of the origin and evolution of theTaurid complex and its individual meteoroid streams.

2. Observations and data reduction

The outburst of Taurid meteors in 2015 was observed globally by nu-merous meteor networks. Here we present the detailed analysis of 33 Tau-rid meteor spectra captured by the spectral All-sky Meteor Orbits System(AMOS-Spec) (Rudawska et al., 2016). Particular focus will be placed on16 of these spectra, which were observed in Modra Observatory, Slovakia.Multi-station observations for 14 of these meteors were provided by four

4

AMOS stations comprising the Slovak Video Meteor Network, which carriesout routine meteor observations on every clear night (Toth et al., 2011; Tothat al., 2015), supplemented by individual observations provided by PavelSpurny (European Fireball Network) and Jakub Koukal (Central EuropeanMEteor NeTwork).

Additional 17 single-station Taurid spectra were observed by AMOS sys-tem in Chile during the testing for two new, now already established, south-ern AMOS stations. The system made observations during the activity peakbetween November 5 and November 10 at San Pedro de Atacama in the Ata-cama Desert. The observational conditions here enable much higher efficiencyin capturing meteor events. These single-station meteors were identified asmembers of the Taurid stream based on their estimated radiant position andgeocentric velocity.

The AMOS-Spec is a semi-automatic remotely controlled video systemfor the detection of meteor spectra. The main display components consist of30mm f/3.5 fish-eye lens, image intensifier (Mullard XX1332), projection lens(Opticon 1.4/19 mm), and digital camera (Imaging Source DMK 51AU02).This setup yields a 100 circular field of view with a resolution of 1600 x1200 pixels and frame rate of 12/s. The incoming light is diffracted by aholographic 1000 grooves/mm grating placed above the lens. The spectralresolution of the system varies due to the geometry of the all-sky lens witha mean value of 1.3 nm/px. A 500 grooves/mm grating was used for theobservations in Chile, providing a mean spectral resolution of 2.5 nm/px.The system covers the whole visual spectrum range from app. 370 nm to 900nm. The spectral response curve of the AMOS-Spec system (camera, imageintensifier, and lens) is given in Figure 1. It was determined by measuringthe known spectrum of Jupiter and is normalized to unity at 480 nm. Thetypical limiting magnitude of the system for meteors is approx. +4 mag.,while only meteors brighter than approx. 0 mag. can be captured alongwith its spectrum. More details about the properties and capabilities of theAMOS systems can be found in Toth et al. (2015). The main disadvantageof the wide-field camera is the interference of the moonlight and bright Moonspectrum causing occasional difficulties in meteor detection. Fortunately, theMoon illumination percentage was descending from 45% to 10% during theactivity peak of the 2015 Taurids (November 4 - November 8) and caused nosignificant problems in spectra analysis.

5

Figure 1: The spectral sensitivity of the AMOS-Spec system. The spectral response curvewas determined by measuring the spectrum of Jupiter, and is normalized to unity at 480nm.

2.1. Spectra reduction

The all-sky geometry of the lenses causes slight curvature of the meteorspectra captured near the edge of the FOV. For this reason, each spectrumwas scanned manually on individual video frames using ImageJ1 program.Before scanning the spectra, all of the analyzed spectral events were correctedfor dark frame, flat-fielded, and had the star background image subtracted.We are particularly interested in the relative intensities of the Na I - 1, Mg I- 2, and Fe I - 15 multiplets, which form the basis of the spectral classifica-tion of meteors established by Borovicka et al. (2005). Each spectrum wasfitted with a simple model accounting for all significant contributions to theobserved spectral profile. The continuum level was fitted by Planck curveat given temperature. For most meteors, the continuum was well fitted byPlanck curve at 4500 K, but in some cases, particularly for fainter meteors,this was lowered down. It should be noted that the fitting procedure doesnot serve as physical interpretation of the spectrum, but is only used for thereduction of contamination caused by the mixture of continuum radiationand weak unrecognizable lines, in order to achieve the best possible fit ofeach spectrum. The spectral lines (low temperature, high temperature, at-mospheric, and wake lines) were fitted using the positions and intensities of

1https://imagej.nih.gov/ij/

6

the most important atomic lines in meteor spectra, as given in Borovicka etal. (2005). The simulated spectrum was fitted by instrumental Gaussian lineprofiles with appropriate full width of typically 5 nm at 1/e of the peak inten-sity. Furthermore, the most important N2 bands of the first positive systempresent in the meteor spectra were fitted using the positions and intensitiestaken from Borovicka et al. (2005), and adjusted for our spectra. The Gaus-sian width of these bands was assumed to be 10 nm. The intensities of allcontributions were adjusted in each spectrum to provide to best possible fit.The intensities of the main meteor multiplets (Mg I - 2, Na I - 1, and Fe I -15) subtracted of the fitted continuous radiation and atmospheric emissions(most prominent lines of O, N, and N2 bands) in the modeled spectrum werethen measured and used in the spectral classification. The scale of each spec-trum was determined using known spectral lines in the calibration spectrumand polynomial fitting of 2nd or higher degree. For most cases the lines ofMg I - 2 (518.2 nm), Na I - 1 (589.2 nm), O I - 1 (777.4 nm), and Fe I - 41(438.4 nm) were used for the scaling of the spectrum, which provided optimalfit for the identification of other lines. Additionally, other prominent spectrallines were employed (e.g., Ca II - 1 at 393.4 nm, O I - 615.3 at nm, or N I- 3 at 744.2 nm.) in cases where other lines were missing in the FOV. Theline identification, correction for spectral sensitivity, reduction of continuumradiation and atmospheric emission, and calculation of relative intensity ra-tios of studied spectral lines are all performed by our own developed Matlabcode. Each spectrum was scanned and scaled on individual frames and thefinal profile was obtained by summing measured intensities on each frame.Only bright enough Taurid spectra with signal-to-noise ratio higher than 5were used for our analysis.

The low resolution of the video meteor spectra allows us only to iden-tify limited number of meteoroid spectral lines compared to high dispersionphotographic spectrographs used mainly in the past (e.g. Ceplecha, 1964;Nagasawa, 1978; Borovicka, 1993). It has been shown that meteor spec-tra consist of two spectral components (Borovicka, 1994). The majority ofthe observed meteoric spectral lines belong to the main, low-temperature (≈4500 K) spectral component, which comprises mostly of the neutral lines ofNa, Mg, Fe, Cr, Ca, and Mn. The high-temperature (≈ 10000 K) spectralcomponent is mostly significant in bright and fast meteors. It consists ofatmospheric O and N lines, and meteoric high-excitation lines of Mg II, Si II,and Ca II. Only three meteoric species (Na I, Mg I, and Fe I) can be reliablyexamined in spectra captured by video spectrographs with resolution similar

7

to our system. Even though other meteoric species are also often observedin meteor spectra (such as the lines of Ca I, Ca II, Si II, Cr I, or Mn I), onlythe multiplets of these three species have sufficient intensity in almost everymeteor spectra to be accurately measured and compared. Nevertheless, thevariations of relative intensities of these lines can reveal the different com-positions of studied meteoroids. The full lists of the most important atomiclines identified in typical low resolution video meteor spectra can be foundin Borovicka et al. (2005) or Vojacek et al. (2015).

2.2. Photometry and astrometry

Previous measurements and orbital analyses of known meteor streamshave demonstrated the astrometric precision of the AMOS system (Toth etal., 2015). The system achieves higher orbital accuracy compared to mostall-sky video systems due to significantly higher pixel resolution. Multiplestation observations of the analyzed spectral events were used to determinethe heliocentric orbital parameters and geocentric trajectory parameters ofthe Taurid meteors. The AMOS systems comprising the SVMN yield a stan-dard astrometric error of 0.03 - 0.05 that translates to the accuracy of tens tohundreds of meters for atmospheric meteor trajectory. The detailed all-skyreduction described by Ceplecha (1987) and Borovicka et al. (1995) was used.Each meteor was measured individually using UFOCapture software for me-teor detection and UFOAnalyzer for astrometric data reduction (SonotaCo,2009).

Magnitudes of meteors were determined by visual calibration based oncomparison with bright stars, planets, and Moon phases in the FOV. Theseapparent magnitudes were then corrected to standard altitude of 100 km atthe observation zenith, and for atmospheric extinction to obtain estimatedabsolute magnitudes. Using this method, magnitude of each studied me-teor was determined from every available observing station, and resultingvalue (Table 5) was obtained as an average from individual station measure-ments. Absolute magnitudes of meteors captured in Chile (Table 2) weredetermined from single-station observations, which is why no error is given.Errors of these measurements from visual calibration are estimated to be ±1magnitude for meteors of -5 magnitude and fainter, and ±2 magnitudes forbrighter meteors. The errors of the estimated photometric masses are influ-enced besides the errors of absolute magnitudes and meteor velocities also bythe effective meteoroid fragmentation model, which was not examined in ourwork. Therefore, the uncertainties of photometric masses is only estimated in

8

given order of magnitude. Few exceptionally bright fireballs among the 2015Taurids caused significant saturation, delivering higher uncertainty in thedetermination of the absolute magnitudes and consequently the photomet-ric masses of the meteoroids.One of our multi-station meteors was observedunder very low convergence angle, which translated into the low accuracyof the obtained orbital elements. These parameters were therefore omittedin our results (Table 5). In future, we plan to apply a new method basedon detailed comparisons to the brightness of the Moon in various phases togain higher accuracy of the estimated magnitudes. Furthermore, to gain thehighest possible accuracy of the determined photometric masses, effectivefragmentation model should be taken into account.

3. Orbital classification

The orbital properties of the 14 Taurid meteoroids observed by multiplestations are in Table 1. The Tisserand’s parameter with respect to Jupitercan be used to distinguish between the different orbit types (cometary, as-teroidal) related to their source regions. It is defined as

TJ =aJa

+ 2

√( aaJ

)(1 − e2) cos i, (1)

where aJ is the semi-major axis of Jupiter, and a, e, i are the semi-majoraxis, eccentricity and inclination of the meteoroid orbit respectively. TheTisserand’s parameter places our sample of meteoroids on the borderline be-tween asteroidal orbits (TJ > 3) and Jupiter-family type cometary orbits(2 < TJ < 3), as would be expected for a stream originating in the short-period comet 2P/Encke (TJ = 3.026). We identified four Jupiter-family typecometary orbits and ten asteroidal orbits among the studied meteoroids.Our sample includes particles from the anticipated outburst caused by the7:2 resonance with Jupiter, along with the background activity of the Tauridstream. Based on the later discussed meteoroid properties (Table 5), we donot assume that found differences between meteoroids defined on asteroidaland Jupiter-family type orbits directly reflect different origin of these parti-cles, but rather demonstrate the broad spatial structure of the stream, whichis defined close to the borderline between these orbit classes. The determinedTaurid orbits are displayed in Figure 2.

For further information about the orbital origin of our sample, we focusedon using the Southworth-Hawkins DSH orbital similarity criterion (South-

9

Table 1: Orbital properties of the multi-station Taurid meteor spectra observed during theoutburst in November 2015. Each Taurid is designated with meteor number, meteor IDbased on the observational date and time, shower assignment, corresponding Southworth-Hawkins criterion (DSH) with respect to the assigned shower, geocentric velocity (vg),semi-major axis (a), perihelion distance (q), eccentricity (e), inclination (i), argument ofperihelion (ω), longitude of the ascending node (Ω), Tisserand’s parameter with respectto Jupiter (TJ), and corresponding orbit type based on TJ (AST for asteroidal orbits,JUP for Jupiter-family type orbits). Meteor no. 7 was omitted due to high uncertaintyof the determined orbital elements, caused by the low convergence angle between the twoobservational stations. Meteor no. 15, and meteors no. 17 - 33 captured in Chile wereonly observed by one station, orbital properties are not available.

No. Meteor ID Shower DSH vg a q e i ω Ω TJ type

1 M20151102 020949 STA 0.08 28.74 1.87 0.316 0.831 5.5 120.19 39.10 3.45 AST± 0.73 0.12 0.008 0.015 0.1 0.25 0.19

2 M20151102 024553 NTA 0.14 27.62 2.04 0.365 0.821 3.3 293.69 219.16 3.27 AST± 0.11 0.03 0.002 0.003 0.1 0.31 0.04

3 M20151103 212219 STA 0.04 27.98 1.89 0.340 0.820 4.7 117.32 40.90 3.45 AST± 0.35 0.06 0.004 0.008 0.1 0.03 0.09

4 M20151103 212454 STA 0.10 29.27 1.90 0.305 0.839 6.6 121.13 40.90 3.40 AST± 0.06 0.01 0.001 0.001 0.1 0.01 0.02

5 M20151105 205304 NTA 0.05 30.12 2.35 0.314 0.866 1.8 298.08 222.95 2.89 JUP± 0.32 0.09 0.003 0.007 0.1 0.03 0.09

6 M20151105 215813 STA 0.03 28.98 2.39 0.354 0.852 4.3 113.45 42.92 2.89 JUP± 0.12 0.05 0.003 0.003 0.1 0.35 0.04

8 M20151105 220251 STA 0.03 28.34 2.17 0.356 0.836 3.7 114.04 42.92 3.11 AST± 0.17 0.07 0.005 0.004 0.7 0.72 0.07

9 M20151105 231200 STA 0.03 27.94 2.02 0.356 0.824 4.9 114.78 42.98 3.28 AST± 0.38 0.08 0.004 0.009 0.1 0.06 0.10

10 M20151105 235959 STA 0.17 26.59 2.55 0.438 0.828 0.8 103.60 42.92 2.83 JUP± 0.20 0.07 0.002 0.005 0.1 0.04 0.05

11 M20151108 234416 STA 0.03 27.27 2.09 0.385 0.816 5.5 111.17 46.01 3.22 AST± 0.17 0.07 0.007 0.005 1.0 1.02 0.08

12 M20151109 041942 NTA 0.05 29.30 2.21 0.330 0.851 1.0 296.74 226.32 3.04 AST± 0.14 0.03 0.002 0.003 0.1 0.06 0.04

13 M20151111 220940 STA 0.04 27.71 2.45 0.400 0.837 5.1 108.00 48.96 2.88 JUP± 0.15 0.05 0.002 0.004 0.1 0.01 0.04

14 M20151116 193459 STA 0.10 25.51 2.08 0.441 0.788 4.9 104.98 53.88 3.28 AST± 0.04 0.01 0.000 0.001 0.1 0.01 0.01

16 M20151123 012004 NTA 0.10 26.06 2.08 0.422 0.797 3.0 286.88 240.24 3.27 AST± 0.03 0.01 0.001 0.001 0.1 0.20 0.02

10

Figure 2: Heliocentric orbits of 14 multi-station Taurid meteors based on orbital elementsgiven in Table 1. Portrayed 4 Northern Taurid orbits (black), 10 Southern Taurids (grey),complemented with the orbit of Earth and Jupiter, and the direction towards the vernalequinox.

worth & Hawkins, 1963) to look for associations of individual meteoroidswith different filaments of the broad Taurid stream. Ten of the studied me-teoroids were assigned to the southern branch of the stream, while four wereidentified as the Northern Taurids. Corresponding DSH values with respectto the assigned Taurid filaments (defined by orbital elements from Porubcanet al., 2006) are given in Table 1. Moreover, we also examined the orbitalsimilarity between the studied meteoroids and the Taurid complex associatedNEOs. Applying similarity threshold value of DSH ≤ 0.10, we found possiblerelation to 10 asteroids: 2005 UR (associated with meteoroids no. 1, 4), 2015TX24 (no. 1, 3, 4, 6, 8, 9), 2003 UV11 (no. 3, 6, 8, 9, 11, 13), 2007 UL12(no. 3, 8, 9, 11, 13, 14), 2010 TU149 (no. 2, 6, 8, 10), 2003 WP21 (no. 14),2010 VN139 (no. 14), 2004 TG10 (no. 5, 12), 2012 UR158 (no. 5, 12), and2014 NK52 (no. 5, 12, 16). Compared orbital elements of the associatedNEOs were taken from the Asteroid Orbital Elements Database currated byTed Bowell and Bruce Koehn. The most associations in the southern branchof our sample was found with asteroids 2015 TX24, 2003 UV11, and 2007UL12. Meteoroids from the northern branch showed most markable similar-

11

ities with asteroid 2014 NK 52. These results support the measurements ofSpurny et al. (2016), who suggested that the 2015 outburst may have beencaused by Taurid filaments associated with asteroids 2015 TX24 and 2005UR. Two of the southern Taurid meteoroids (no. 1 and no. 4) originatedon orbits closely resembling the orbit of asteroid 2005 UR (DSH = 0.06 and0.07). The orbital similarity of 2005 UR with two bright Southern Tauridfireballs observed over Poland was also noted by Olech et al. (2016). Theobtained orbital similarities with aforementioned asteroids are however stillrather indefinite and could also be, given the broad structure of the Tauridcomplex, only incidental. In this work, we will focus on using the spectraland physical properties of Taurid meteoroids to look for features suggestinggeneric associations with asteroidal origin.

4. Spectral classification

As mentioned earlier, the low resolution of video spectrographs only al-lows us to study limited number of spectral lines in meteor spectra. However,the relative intensities of the three main meteoric species (Na, Mg, Fe) ob-served in the visual spectrum can reveal the rough composition of meteoroids.This is the basis of the spectral classification established by Borovicka et al.(2005), which distinguishes different types of bodies based on the intensitiesof spectral multiplets Na I - 1 (589.2 nm), Mg I - 2 (518.2 nm), and Fe I- 15 (526.0 - 545.0 nm). The modeled contributions of all recognized linesof the Fe I - 15 multiplet were summed. These multiplets were chosen asthey can be well measured in most meteor spectra (see Section 2.1), andall lay in the region of high instrumental spectral sensitivity. Nevertheless,the measurement of these lines must be careful, because the Na multipletoverlaps with a sequence of N2 bands, while the wake lines of Fe can influ-ence the intensity of Mg line. All these effects were accounted for in thereduction process described in Section 2.1. As the spectral sensitivity of theAMOS-Spec is in this region similar to the sensitivity of the system used byBorovicka et al. (2005), we are able to apply the same classification also forour observations. It must be noted that the classification of Borovicka et al.(2005) was originally developed for fainter meteors in the magnitude range+3 to -1, corresponding to meteoroid sizes 1 - 10 mm. Our system observesmeteor spectra of -1 to -10 magnitude, corresponding to meteoroid sizes ofapp. few mm to tens of cm. However, we expect that the same physicalconditions fitted by the thermal equilibrium model applied for bright photo-

12

Figure 3: The spectral classification of 33 Taurid meteors observed in Slovakia and Chilein November 2015. Diagram shows the measured relative intensities of multiplets Mg I -2, Na I - 1, and Fe I - 15. Meteoroids are marked with numbers ordered according to theobservational date. Meteors 1 - 16 were observed in Slovakia, while meteors 17 - 33 werecaptured in lower spectral resolution in Chile.

graphic meteors (Borovicka, 1993) as well as for fainter meteors (Borovickaet al., 2005), can be also assumed for the population observed by our system.The effect of self-absorption in brighter meteor spectra was examined in in-dividual cases. The instrumental pixel saturation in spectra only occurred inindividual meteors during bright flares and posed no problem for most me-teors in the resolution provided by the 1000 grooves/mm grating. Individualframes with saturated meteor spectra were excluded from the summation forfinal spectral profiles and calculation of relative line intensities.

The spectral classification of Taurid meteoroids observed in Slovakia andChile is on Figure 3. The majority (24) of Taurid meteoroids represent normaltype spectral class with similar ratios of Na I, Mg I, and Fe I intensities.

13

However, we observe apparent variations in Fe and Na content among themeteoroids in our sample. Particularly, the studied spectra revealed largedispersion of iron content in individual cases, ranging from Fe-poor bodiesto almost chondritic Fe/Mg ratios (Figure 4). Normal type meteoroids arepositioned in the middle of the ternary diagram, with content close to theratios characteristic for chondritic composition. It was noted by Borovicka etal. (2005) and later confirmed by Vojacek et al. (2015) that the majority ofobserved meteoroids fall in the ternary diagram below the so called chondriticcurve, which represents the expected range for chondritic composition, dueto the lower content of iron. This is also the case for our sample of Tauridmeteoroids, though two of the measured spectra showed quite significant Feintensities (meteoroids no. 5 and 19). Furthermore, we also identified 8Taurids depleted in iron, representing Fe-poor class, and two Na-enhancedTaurid meteoroid (no. 13 and 16). As can be clearly seen in Figure 3, allof the Taurids are positioned right from the center of the ternary diagram,exposing slight Na line intensity enhancement over Mg in all of the studiedmeteoroids. This supports the results of Borovicka et al. (2005), who foundNa-enhancement in three of the 6 observed Taurid spectra.

The Na line intensity is certainly one of the most interesting features inmeteor spectra, revealing the thermal history of meteoroids. The presence ofsodium in meteor spectra can be used as a tracer of volatile phases associatedwith cometary origin. During the formation of the solar system, volatile ele-ments were depleted from the inner protoplanetary disk due to intense solarradiation (Despois, 1992). We expect higher Na abundance in unprocessedbodies such as comets. Borovicka et al. (2005) pointed out three popula-tions of Na-free meteoroids: iron bodies on asteroidal orbits; meteoroids withsmall perihelia (q ≤ 0.2 AU), where Na was lost by thermal desorption; andNa-free meteoroids on Halley type orbits, where Na loss was possibly causedby irradiation of cometary surfaces by cosmic rays in the Oort cloud. Thedepletion in Na for small meteoroids is observed as a function of the timeexposure to solar radiation, which predicts that asteroidal particles will beNa-poor materials (Trigo-Rodrıguez & Llorca, 2007). Even though all ofour Taurid spectra have shown unambiguous sodium content, we observe ap-parent effect of Na depletion as a function of perihelion distance (Figure 5).The observed dependence of Na/Mg ratio on perihelion distance however canbe partially influenced by the effect of saturation/self-absorption in bright(larger) meteors, while the measurement of line intensity in fainter spectra(typical for smaller particles) is a subject of higher uncertainty. The errors

14

Figure 4: Spectral profiles of Taurid meteoroids no. 5, 7, and 11. The spectra show similarfeatures with significant dispersion of the Fe I - 15 multiplet intensity from the highest(no. 5) to the lowest (no. 11). Displayed spectral profiles are not corrected for the spectralsensitivity of the system (Fig. 1).

15

Figure 5: The dependence of the Na/Mg line intensity ratio on perihelion distance. Thevarying sizes of the marks display photometric masses of individual meteoroids in loga-rithmic scale.

of determined Na/Mg and Fe/Mg intensity ratios from signal-to-noise ratioin individual observations are given in Table 5. As can be seen in Figure 5,the Na enhancement was preferred in smaller Taurid meteoroids, in whichthe Na/Mg ratio was determined with higher uncertainty. Furthermore, me-teoroids with high surface-area/volume ratios are expected to suffer higherdepletion of Na and other volatiles through solar radiation (Trigo-Rodrıguez& Llorca, 2007).

The increased values of Na/Mg ratios in Taurid spectra are not too sur-prising. It was pointed out that the Na/Mg line intensity ratio varies widelyin cometary meteoroids (e.g. Borovicka et al., 2005; Rudawska et al., 2016),and is observed as a function of meteor velocity. Sodium is characterizedby a low ionization potential, and the Na line is of low excitation (2.1 eV)compared to the Mg line (5.1 eV) (Borovicka et al., 2005). Consequently, weexpect Na line intensity to be dominant at lowest velocities (temperatures)and lower at highest velocities (temperatures). This effect is only observedfor meteor velocities below 40 km/s and increases for velocities below 20km/s. Typical Taurid velocities of approx. 28 km/s would therefore slightlyfavor Na intensity over Mg, particularly for fainter meteors. Bright fireballscaused by more massive meteoroids can reach very high temperatures, atwhich both Na and Mg become fully ionized, as observed in Figure 5.

Moreover, also the intensities of atmospheric emission lines of O I andN I, as well as the N2 bands increase with the meteor velocity. As a resultof the moderate velocities of the Taurid meteors, the atmospheric emission

16

lines were clearly present in the spectra (Figure 4), but did not achieve highintensities, allowing us to readily subtract them from the spectral profilesbefore determining the intensities of the main meteoroid multiplets.

The brightest multi-station Taurid in our sample was observed on Novem-ber 5 at 23:21:00 UT. This fireball (marked no. 9 in Figure 3) of estimated-8.4 absolute magnitude exhibited bright flare, which enabled us to capturetwo orders of its spectrum. The 1st order and the 2nd order spectral profilesare on Figure 6. While the 1st order spectrum unveils more detailed featureswith emission lines of the high-temperature spectral component, the satu-ration in several frames caused by the flare restricts us from analyzing therelative intensities of the main spectral multiplets throughout the entire me-teor flight. For this reason, we used the higher dispersion 2nd order spectrumto determine the spectral classification of this Taurid. Clearly, the relativeintensities of Mg I - 2, Na I - 1, and Fe I - 15 differ in the two spectral profiles.We can observe the lines of Fe I - 15 blending into the Mg I - I radiation andincreasing its intensity. This demonstrates the uncertainty of determiningrelative intensities in bright spectra of meteor flares.

Example of a typical lower-resolution Taurid spectrum captured by a sys-tem in Chile using the 500 grooves/mm grating is on Figure 7. Even thoughwe are not able to study fainter spectral features in these profiles, the intensi-ties of the main spectral multiplets are well measurable. The blending of MgI - 2 and Fe I - 15 multiplets is more apparent in low-resolution spectra (Fig-ure 7), but it was resolved for the relative intensity measurement by fittingthe resulting profile shape with intensities of all expected spectral lines ingiven area. The general results are in good agreement with the spectral clas-sification of Taurids based on observations in Slovakia. Estimated absolutemagnitudes of studied single-station Taurid meteors captured with spectrain San Pedro de Atacama Chile are in Table 2.

Generally, the spectral characteristics of the measured Taurid spectra,which are defined by mostly normal-type composition with Fe content belowthe chondritic values and increased Na content in most meteors point towardsthe cometary origin of all of the studied Taurid meteoroids. Nevertheless, weobserve large variations of Fe and notable Na dispersion in Taurid spectra,suggesting Taurids are very heterogeneous population of meteoroids. Twoof the Taurid spectra exhibited slightly enhanced Fe intensities (no. 5 andno. 19). The spectrum of meteor no. 19 was captured from a short, brightflare (estimated -9.4 absolute magnitude) and in lower spectral resolution,which brings higher uncertainty of measured relative intensities related to

17

Figure 6: Brightest multi-station Taurid meteor (no. 9) in our sample of estimated -8.4 magnitude captured on November 5 at 23:21:00 UT, and the spectral profiles of thefirst order spectrum (upper picture) and second order spectrum (lower picture) showingdifferent relative intensities of the Mg I - I, Na I - 2, and Fe - 15. multiplets. Displayedspectral profiles are not corrected for the spectral sensitivity of the system (Fig. 1).

18

Figure 7: Example of a low-resolution Taurid spectrum (no. 29) captured on November8 at 07:00:43 UT in Chile. Displayed spectral profile is not corrected for the spectralsensitivity of the system (Fig. 1).

possible effects of saturation/self-absorption. The observed heterogeneity isnot unexpected for a stream of such remarkable spatial dispersion. In ad-dition to the original heterogeneity of the parent body of the stream, thediverse thermal evolution in different parts of the stream would result in theobserved compositional variations. We did not find any spectral features,which would clearly indicate asteroidal origin in any of the meteoroids in oursample, though the relative intensities Na I, Mg I, and Fe I in several mete-oroids were not too distant from the values expected for chondritic bodies.For more information, we focused on using the multi-station observations of14 Taurid meteors to determine the physical characteristics of the studiedmeteoroids.

5. Physical properties of Taurid meteoroids

The primary focus of this section will be placed on the material strengthsof Taurid meteoroids, which could indicate additional asteroidal source ofmeteoroids in the stream, and also tell us about the potential of this streamin producing meteorites.

The first physical parameter we need to determine is the meteoroid mass,which is also later used in the calculation of other physical parameters. Theactivity of the Taurid meteor shower is characteristically rich in bright fire-balls indicating larger particles than what we observe in most cometary show-

19

Table 2: Estimated absolute magnitudes (Mag) of single-station Taurid meteors capturedwith spectra in San Pedro de Atacama, Chile. Each meteor is designated with meteornumber and meteor ID based on the observational date and time. Errors of absolutemagnitudes are estimated to be ±1 magnitude for meteors of -5 magnitude and fainter,and ±2 magnitudes for brighter meteors.

No. Meteor ID Mag

17 M20151106 042700 -5.518 M20151106 052449 -4.519 M20151106 062854 -9.420 M20151106 063233 -3.521 M20151106 070343 -8.622 M20151107 031208 -3.923 M20151107 061350 -7.724 M20151107 063741 -7.825 M20151107 072801 -5.526 M20151107 073808 -5.227 M20151108 020459 -2.228 M20151108 025443 -11.629 M20151108 070043 -10.430 M20151108 071904 -2.831 M20151108 083951 -4.932 M20151110 004036 -6.833 M20151110 054715 -4.1

20

ers (Wetherill, 1974). It was suggested that Taurids contain very large me-teoroids at least in the hundreds of kilograms range, which is more than anorder of magnitude larger than other showers (Brown et al., 2013; Ceplechaet al., 1998). Our sample included Taurid meteors ranging from -1.3 to -8.4 magnitude, only confirming these characteristics. The meteoroid masseswere determined photometrically, i.e. by integrating along the light curvesof meteors. The standard luminosity equation (e.g. Bronshten, 1983) can berewritten in the form:

m∞ = 2

∫ tE

tB

Idt

τlv2. (2)

Here, m∞ is the original mass of the meteoroid, v is the meteor velocity, and Iis the light intensity of the meteoroid, integrated in the time interval from thebeginning of the light curve (tB) to the end (tE). The luminosity efficiencyfactor, τl, represents the fraction of the meteoroid’s kinetic energy, which hasbeen converted into luminosity in the visual range. For the calculation of τl,we applied the approach described by Hill et al. (2005). The method usesa relation between the luminous efficiency factor and universal excitationcoefficient ζ defined by Jones & Halliday (2001):

τl = 2( εµ

) ζv2, (3)

where the values of ζ are split into several regimes based on the meteorvelocity and the ratio of the mean excitation energy to the atomic mass(ε/µ) is computed for meteoric atoms and ions producing the most radiation(e.g. Ceplecha et al., 1998), assuming the majority of the observed radiationcomes from the meteoroid atoms (Bronshten, 1983). The estimated massesof the meteoroids in our sample range from 1 to 1500 grams (Table 5).

5.1. Material strength

Meteor observations can be used to characterize the material strengthof meteoroids by determining empirical parameters KB and PE, which arefunctions of the beginning and the terminal height of the meteor luminouspath. Fragile particles tend to crumble under lower pressures resulting inhigher beginning height of the observed luminous trajectory. On the contrary,only stronger materials can withstand the pressure and reach lower parts ofour atmosphere, and only the most robust particles will eventually fall onthe Earth’s surface. It is observed that depending on the entry velocity and

21

meteoroid mass, the ablation begins at higher heights for cometary bodiesthan asteroidal ones (Ceplecha & McCrosky, 1976; Koten et al., 2004).

The KB and PE parameters defined by Ceplecha (1968), and Ceplecha &McCrosky (1976), and later revised by Ceplecha (1988) can be expressed as:

KB = log ρB + 2.5 log v∞ − 0.5 log zR, (4)

PE = log ρE − 0.42 logm∞ + 1.49 log v∞ − 1.29 log cos zR. (5)

Here, ρB and ρE are the atmospheric densities at beginning and end height ofthe luminous trajectory, v∞ is the pre-atmospheric meteor velocity, zr is thezenith distance of the radiant, and m∞ is the original photometric mass cal-culated using luminous efficiency factor in accordance with Ceplecha & Mc-Crosky (1976). The values of atmospheric densities at different heights wereobtained using the MSIS-E-90 Atmosphere Model of Hedin (1991). Theseparameters form a basis for the empirical classification of meteoroid mate-rial strength (Table 3). The KB classification differentiates between fragilecometary bodies (group D) characteristic e.g. for Draconid meteors, typicalcometary bodies (group C), dense cometary bodies (group B), carbonaceous-type bodies (group A), and strong meteoroids of asteroidal origin (group ast).Similarly, the classification based on the PE parameter distinguishes betweencometary bodies (group III), carbonaceous-type bodies (group II), and as-teroidal bodies (group I).

We clearly observe a large scatter of material strengths among Tauridmeteoroids in our sample (Table 5). Following the KB classification (Fig-ure 8), the majority of our Taurids fall into the C group characteristic forstandard cometary bodies, with only two meteoroids appearing to be morefragile (D group) and two meteoroids of slightly higher strength (B group).It should be noted that the KB parameter (Eq. 4) assumes that beginningheight does not depend on meteoroid mass (meteor magnitude). While thisis true for a narrow range of meteor magnitudes, Koten et al. (2004) haveshown that the beginning height increases with mass for most meteor show-ers. The classification of the two brightest Taurids in our sample as type D isprobably consequence of this fact. The obtained KB values are neverthelesswhat we would expect from cometary bodies originating in 2P/Encke. How-ever, we see significantly higher dispersion of Taurid material strengths inthe PE classification (Figure 8). Here, only three meteoroids have PE valuescharacteristic for cometary bodies, while the majority is defined as group II,similar to carbonaceous bodies, and four Taurids even exposed end-height

22

Table 3: Meteoroid material types based on material strength parameters KB and PE , asdefined by Ceplecha (1988). The differences between groups C3 and C2, IIIAi and IIIAare based on orbital properties (for further information, see Ceplecha, 1988)

Material type KB group KB PE group PE

Fragile cometary D KB < 6.6 IIIB PE ≤ -5.70

Regular cometarylong-period

C3 6.6 ≤ KB < 7.1 - -

Regular cometarylong-period

C2 6.6 ≤ KB < 7.1 IIIAi -5.70 < PE ≤ -5.25

Regular cometaryshort-period

C1 6.6 ≤ KB < 7.1 IIIA -5.70 < PE ≤ -5.25

Dense cometary B 7.1 ≤ KB < 7.3 - -

Carbonaceous chondrites A 7.3 ≤ KB < 8.0 II -5.25 < PE ≤ -4.60

Ordinary chondritesAsteroids

ast 8 ≤ KB I -4.60 < PE

Figure 8: Classification of meteoroid material strength based on parameters KB and PE

compared to the orbital classification of meteoroids based on Tisserands parameter. Tauridmeteoroids (filled marks) are compared to diverse population of 22 meteoroids taken fromRudawska et al. (2016). Blue marks represent meteoroids on Halley-type cometary orbits,yellow marks designate Jupiter-family cometary orbits, and black represent asteroidalorbits. The varying sizes of the marks display photometric masses of individual meteoroidsin logarithmic scale.

23

strengths typical for rocky asteroidal bodies (group I). Similarly, Brown etal. (2013) found that most Taurids appear to be type II or type III fire-balls. Great variations of material strength in Taurid meteoroids includingvery strong objects among fireballs were previously indicated by Borovicka(2006). It seems that while the beginning heights of Taurids are typicallycometary, the end heights are usually lower and more dispersed (Figure 9).Ceplecha (1988) noted that parent bodies of group II and group A meteoroidsare perhaps partly asteroids and partly comets (Wetherill & Revelle, 1981).However, the discovered Taurid features more likely suggest heterogeneousstructure of its meteoroids, which are of cometary nature but contain solid,possibly carbonaceous inclusions. The diversity of material types observed onTaurid orbits could be explained by the inhomogeneous interior of the parentcomet (2P/Encke or earlier larger body). The existence of solid inclusions,which relate to CI and CM carbonaceous chondrites has also been indicatedin a small population of Perseid meteoroids with high mineralogical densities(Babadzhanov & Kokhirova, 2009), and confirmed in the material of Leonidmeteoroids (Borovicka, 2006). Surprisingly low material strength is observedin meteoroid no. 5, which exhibited the highest content of iron in our sam-ple. While this is the second brightest Taurid (estimated -8.1 magnitude),no signs of significant effects of saturation/self-absorption, which could causeover-estimation of iron intensity, were observed in the spectrum. There is noclear evidence supporting asteroidal origin in any of the Taurid meteoroids,nevertheless, the connection to carbonaceous chondrites remains unrefuted.Furthermore, the diversity of material strengths among Taurids could alsorelate to the thermal history of these meteoroids. The two Na-enhanced me-teoroids in our sample (No. 13 and No. 16) were both identified as groupI meteoroids (Table 5). It is usually observed that meteoroids depleted insodium are stronger than other particles (Borovicka et al., 2005). Exposureto stronger solar radiation in lower perihelia orbits may have also causedslight structural disruption, which resulted in the lower material strength inKB classification. The dependence of material strength parameter KB onsodium content (Figure 10) is apparent in the majority of the sample, butwas not manifested in a branch of four meteoroids, suggesting this effect hasonly partial influence on the observed material strength. The observed effectcould be also partially influenced by the enhancement of Na/Mg in brightermeteors, possibly affected by saturation/self-absorption, as was mentionedearlier. The overall physical properties displayed in Table 5 indicate fourmeteoroids of group I, also characterized with higher mineralogical densi-

24

Figure 9: Comparison of the beginning and terminal heights of meteors. Taurid meteoroids(filled marks) are compared to diverse population of 22 meteoroids taken from Rudawskaet al. (2016). The marking of meteoroids is the same as in Figure 8.

ties, as the best possible candidates for non-cometary origin, however, theasteroidal source cannot be confirmed in any of the studied meteoroids.

5.2. Mineralogical density

Mineralogical density is certainly one of the most important physicalproperties defining the nature meteoroids. There are numerous methods forthe determination of meteoroid densities. Bulk densities of meteoroids canbe calculated from meteor observations by simultaneously fitting the decel-eration and the light curves of meteors (e.g. Borovicka et al., 2007; Kikwayaet al., 2011). The effective fragmentation model must be taken into account.

Mineralogical density represents the density of the substance, irrespectiveof the structure and shape of the body (Bronshten, 1983). The actual (min-eralogical, or grain) density of the meteoroid material differs from its bulk(volume) density due to the porous structure, presence of voids and porosi-ties, which are characteristic for most meteoroids. We applied the methodsuggested by Ceplecha (1958) and developed by Benyukh (1968), which usesthe heat conductivity equation to obtain the mineralogical density of mete-

25

Figure 10: The dependence of the Na/Mg line intensity ratio on material strength pa-rameter KB . The varying sizes of the marks display photometric masses of individualmeteoroids in logarithmic scale.

oroids. The equation of heat conductivity can be written in the form:

2TB(λδmc)1/2

Λ=

v5/2∞ ρ

(b cos zr)1/2, (6)

where TB is the temperature of the frontal surface of a meteoroid; λ is themeteoroid heat conductivity; δm is the mineralogical density of the meteoroid;c is the specific heat of the meteoroid; Λ is the heat transfer coefficient in thebeginning of evaporation; v∞ is the pre-atmospheric velocity of the meteoroid;ρ is the atmospheric density; zR is the zenith distance of the radiant; andb = 1/H is the air density gradient, where H is the scale height. Theatmospheric parameters (densities, temperatures) were determined from theMSIS-E-90 Atmosphere Model of Hedin (1991). Values of parameters in Eq.6 are applied at the meteor beginning height.

The laboratory data of λ, δm, and c are available for various types of rocks,minerals, and metals (Berch et al., 1949), applied values are given in Table 4.The values of TB and Λ can be assumed for meteoric stone and iron particles.According to Levin (1956), the temperature of a meteoroid frontal surfaceat the beginning height equals 1600 K for friable stone particles, 2400 K fordense particles, and 2800 K for iron particles. The heat transfer coefficientΛ is assumed to be 1 for stone meteoric particles and 0.75 for iron bodies(Levin, 1956). By applying these values and defining the decimal logarithm

26

Table 4: Laboratory data of mineralogical density (δm), specific heat (c), heat conductivity(λ), and corresponding function f(δm) (Eq. 7) for different materials (Berch et al., 1949).

Material δm c x 107 λ f(δm)(g cm-3) (erg g-1 deg-1) (erg s-1 cm-1 deg-1)

Brick (Kizel’gur) 0.40 0.75 9.3 x 103 8.72Sand (river) 1.52 0.70 2.7 x 104 9.39Sandstone 2.45 0.93 1.7 x 105 9.93Schist 2.67 0.71 1.5 x 106 10.13Granite 2.70 0.65 2.5 x 105 10.15Basalt 2.90 0.85 2.0 x 105 10.20Fluorit 3.20 0.85 1.0 x 106 10.25Fyalit 4.40 0.55 2.0 x 106 10.52Gematit 5.26 0.61 2.4 x 105 10.66Arsenopirit 6.07 0.43 3.8 x 106 10.82Iron 7.60 0.44 7.3 x 106 11.00

27

Figure 11: The dependence of the function f(δm) ) (Eq. 7) on mineralogical density δm.The points represent values for different materials assuming physical parameters, whichwere used by Benyukh (1968).

of the left-hand side of Eq. 6 as the function

f(δm) = log[2TB(λδmc)

1/2

Λ

], (7)

we are able to construct the dependence of f(δm) on δm (Figure 11). Thedecimal logarithm of the right-hand side of Eq. 6,

f(δm) = log[ v

5/2∞ ρ

(b cos zr)1/2

], (8)

can be calculated from precise video observations and the applied atmospheremodel. Using these values of f(δm) in the diagram on Figure 11, we are ableto determine the mineralogical density δm of meteoroids.

This technique was used by Benyukh (1974) to estimate the mineralogi-cal densities of over 3000 meteoroids from a database of meteors captured bySuper-Schmidt cameras. Recently, Babadzhanov & Kokhirova (2009) appliedthe aforementioned method to determine mineralogical densities of 501 me-teoroids observed by photographic cameras, delivering results in satisfactory

28

agreement with the results of Benyukh (1974). Considering the sensitivityof the AMOS system, which is similar to the sensitivity of Super-Schmidtcameras, we expect to observe comparable beginning heights of meteors andthus comparable results for mineralogical densities.

We determined high spread of mineralogical densities in our sample ofTaurids ranging from as low as 1.3 g cm-3 to more expected values of approx-imately 2.5 g cm-3 (Table 5). While over half of the meteoroids are positionedclose the upper limit of this scale, we also observe a non-negligible popula-tion of low density Taurids. It should be taken into account that the appliedtechnique is highly dependent on the observed beginning height of meteors(Figure 12). Figure 12 reveals that the two most massive meteoroids in oursample are the ones with the lowest mineralogical density. Koten et al. (2004)shown that the beginning heights are functions of meteoroid mass and thatthis dependence can exhibit different behavior for different meteor showers.The works of Benyukh (1974) and Babadzhanov & Kokhirova (2009) providedmineralogical densities for a sample of photographic Taurids with the meanvalue of 2.7 ± 0.2 g cm-3. While the heterogeneity of the Taurid stream couldprovide such observed material variations, the results could also be affectedby the slightly higher sensitivity of AMOS system, and the differences in theapplied atmospheric model. Nevertheless, we identified compact populationof Taurids positioned close to the modus of the mineralogical density valuesin our sample at around 2.5 g cm-3. Meteoroids with higher mineralogicaldensity expectedly also show higher values of material strength parametersKB and PE. Whereas the relation between mineralogical density and KB isparticularly apparent due to the common dependence on meteor beginningheight, the PE - density functionality is more complex but still evident.

5.3. Dynamic pressure

The bulk strengths of meteoroids can be also inferred from the loadingram pressure causing their fragmentation. During the aerodynamic loading,the internal stresses in the meteoroid are proportional to the exerted rampressure (Fadeenko, 1967), and so the stagnation pressure at breakup canbe used as a measure of meteoroid strength in the disruption (Popova et al.,2011). The fragmentation strength of a meteoroid can be simply taken asthe dynamic pressure acting at the front surface of the meteoroid:

p = ρhv2h, (9)

29

Figure 12: The dependence between the mineralogical density and beginning height ofTaurid meteoroids. The varying sizes of the marks display photometric masses of individualmeteoroids in logarithmic scale.

where ρh and vh are the atmospheric density and meteor velocity at the heightof the fragmentation. Both of these quantities can be easily measured fromthe observational data and the atmospheric model, but the determination ofthe fragmentation height is often difficult. Various methods can be appliedfor this purpose, depending on the type of available data and the type ofmeteoroid fragmentation (Popova et al., 2011).

The fragmentation heights of the Taurid meteoroids presented here weredetermined from meteor light curves. The meteor flares (sudden transientincreases of brightness) observed mainly in bright fireballs are caused by thesudden release of dust or small fragments from the meteoroid. The positionof the flares in Taurid light curves was used to infer the fragmentation height.However, not all of the analyzed Taurid meteors exhibited recognizable flares,and therefore we were not able to determine reliable values of dynamic pres-sures for these, characteristically fainter cases. The inferred fragmentationstrengths at first breakup of 8 Taurids for which we observed recognizableflares are in Table 5.

Taurid fireballs were previously observed to break-up under 0.05 - 0.18MPa (Konovalova, 2003). The inferred dynamic pressures in our sample ofTaurid meteoroids ranges from 0.02 - 0.10 MPa (Table 5), with the averagefragmentation strength of 0.05 MPa placed on the lower limit of the results ofKonovalova (2003). Expectedly, meteoroids of higher mineralogical density

30

fragmented under higher pressures than the low-density bodies. The major-ity of C type Taurids were characterized by low fragmentation strengths withvery low values observed in the two D type meteors (0.024 and 0.034 MPa).While there are some discrepancies observed between the KB and PE classi-fication with respect to determined dynamic pressures, the obtained valuesare in general characteristic for fragile cometary bodies. Tensile strengthsof average small shower meteoroids were examined by Trigo-Rodrıguez &Llorca (2006), who found that next to Geminids, the Taurids contained thestrongest shower meteoroids, with bulk densities in some fireballs compa-rable to the Tagish Lake meteorite (Hildebrand et al., 2006). In a surveyof instrumentally observed meteorite falls, Popova et al. (2011) found thatthese bodies show very low bulk strengths of 0.1 to 1 MPa on first breakup.The lower limits of this range are characteristic for the weakest observedmeteorite falls - carbonaceous chondrites Tagish Lake, Maribo and Sutter’sMill (Borovicka et al., 2015). Based on the inferred dynamic pressures inour sample, we can assume that Taurid meteoroids are on average weakercompared carbonaceous chondrites. Still, the stream could include particlesof comparable mechanical strengths, as the upper limits of the determinedfragmentation pressures reach 0.1 MPa. The three instrumentally observedcarbonaceous meteorite falls have shown several common characteristics. Allobjects were initially very large in size and showed features indicating veryfragile structure (early fragmentation, high end heights). The very similarorbits of Maribo and Sutter’s Mill suggest common and relatively recent ori-gin (Jenniskens et al., 2012). The entry velocities of these meteorite falls arearound 28 km/s, which is also the average speed of Taurid meteors, and veryclose to the cut-off velocity suggested by modeling for production. While theTaurids remain the most promising major meteor shower candidate for mete-orite recovery (Brown et al., 2013; Madiedo et al., 2014; Brown et al., 2016),the low strengths of its meteoroids, though in the upper limits comparableto the carbonaceous chondrites, would require significant initial mass of theimpacting body.

6. Conclusion

We presented an analysis of spectral and physical properties of Tauridmeteoroids observed during the outburst in November 2015. In a sampleof 14 multi-station Taurid meteors with spectra, we identified 10 SouthernTaurids and 4 meteors belonging to the Northern Taurids shower. The Tis-

31

Table 5: Physical properties of the studied multi-station Taurid meteoroids. Each Tauridis designated with meteor number, absolute magnitude (Mag), beginning height (HB),terminal height (HE), Na/Mg intensity ratio, Fe/Mg intensity ratio, spectral class, ma-terial strength parameters (KB , PE) and corresponding material type, dynamic pressure(p) and mineralogical density (δm). The uncertainties of photometeric masses are in thegiven order of magnitude; the accuracy of the beginning and terminal heights is on theorder of the last digit. Meteor IDs are the same as in Table 1.

No. Mag Mp(g) HB(km) HE(km) Na/Mg Fe/Mg Class KB PE p(MPa) δm(g cm−3)

1 -4.7 40 104.0 70.1 1.74 0.65 Fe poor 6.76 C1 -5.32 IIIA - 1.94± 0.7 0.14 0.08 0.04 0.15 0.18

2 -4.5 60 99.2 52.7 1.59 1.18 normal 7.11 B -4.36 I - 2.50± 0.7 0.16 0.13 0.01 0.13 0.05

3 -5.1 50 99.1 69.3 1.30 0.77 normal 7.05 C1 -5.35 IIIA 0.063 2.43± 0.8 0.14 0.10 0.02 0.15 0.004 0.15

4 -6.0 150 99.7 60.7 1.17 1.05 normal 7.05 C1 -4.98 II 0.099 2.44± 1.2 0.09 0.09 0.01 0.21 0.003 0.10

5 -8.1 1000 108.8 56.6 1.23 1.73 normal 6.34 D -5.13 II 0.034 1.32± 2.0 0.05 0.06 0.02 0.36 0.002 0.15

6 -2.1 1 99.3 72.2 2.61 1.70 normal 7.09 C1 -5.00 II - 2.48± 0.7 0.50 0.37 0.01 0.11 0.07

7 -5.6 40 - - 1.48 1.15 normal - - - - - -± 2.0 0.07 0.06

8 -1.8 1 99.2 73.4 1.73 0.72 normal 7.07 C1 -5.04 II - 2.47± 0.7 0.35 0.22 0.02 0.13 0.13

9 -8.4 1500 106.0 60.5 1.59 1.10 normal 6.53 D -5.51 IIIA 0.024 1.58± 1.4 0.07 0.06 0.03 0.35 0.002 0.15

10 -1.9 1 99.3 68.5 2.39 0.61 Fe poor 7.04 C1 -4.90 II - 2.41± 0.6 0.22 0.10 0.02 0.11 0.13

11 -5.0 70 104.9 63.7 1.49 0.61 Fe poor 6.61 C1 -5.14 II 0.018 1.70± 0.9 0.06 0.04 0.02 0.16 0.002 0.08

12 -5.5 150 104.4 68.1 1.57 0.80 normal 6.74 C1 -5.17 II 0.027 1.93± 0.6 0.09 0.06 0.01 0.11 0.002 0.13

13 -1.3 1 103.0 62.8 2.90 0.60 Na enhanced 6.74 C1 -4.46 I - 1.92± 0.9 0.38 0.16 0.02 0.16 0.08

14 -3.9 50 99.8 54.5 2.12 0.64 Fe poor 6.93 C1 -4.42 I 0.054 2.23± 0.9 0.09 0.05 0.02 0.14 0.003 0.02

16 -2.9 10 98.1 57.0 2.44 0.39 Na enhanced 7.15 B -4.49 I 0.081 2.54± 0.7 0.23 0.09 0.01 0.11 0.002 0.03

32

serand’s parameter defined the orbital origin of the sample on the borderlinebetween asteroidal orbits (10 meteors) and Jupiter-family type cometary or-bits (4 meteors), as would be expected for a stream originating in the short-period comet 2P/Encke. Furthermore, we used a simple method based onthe Southworth-Hawkins DSH orbital similarity criterion to look for associ-ations with related Taurid complex NEOs. We found possible association to10 asteroids satisfying the similarity threshold value of DSH ≤ 0.10. Themost meteor associations in the southern branch of our sample was foundwith asteroids 2015 TX24, 2003 UV11, and 2007 UL12. Meteoroids from thenorthern branch showed most markable similarities with asteroid 2014 NK52.However, we argue that based on the determined spectral and physical prop-erties of Taurid meteoroids, these associations may be purely incidental.

The spectral analysis revealed large dispersion of iron content in Tauridmeteoroids, ranging from Fe-poor bodies to meteoroids with almost chon-dritic Fe/Mg ratios. Na content also varied slightly and was found to bea function of the meteoroid perihelion distance, assigning Na loss to spaceweathering processes, particularly the thermal desorption. It was also foundthat the Na enrichment is preferred in smaller particles. Nevertheless, theincreased Na line intensity over Mg was found in all of the studied spectra,indicating significant volatile content in Taurid meteoroids. The overall spec-tral classification of Taurid meteors observed in Slovakia and Chile showedcorresponding results dominated by normal-type spectral classes and severalcases of Fe-poor and Na-enhanced meteoroids. Besides the almost chondriticcontent of iron in few of the studied cases, there are no spectral featuresclearly suggesting asteroidal origin in any of the analyzed meteoroids.

Overall, the determined material strengths showed typically cometarycharacteristics in the KB classification, dominated by C group meteoroids,and much more dispersed values in the PE classification, in which the major-ity of Taurid meteoroids fall in the group II, characteristic for carbonaceousbodies. The discovered features suggest that Taurids are heterogeneous pop-ulation of meteoroids, which are cometary in nature but contain solid, possi-bly carbonaceous inclusions. The diversity of the observed material may becaused by the reprocessing of meteoroid surfaces, but possibly also reflects theauthentic inhomogeneity of 2P/Encke or earlier parent comet of the stream.Four meteoroids of group I characterized with higher mineralogical densitieswere identified as the best possible candidates for non-cometary origin, how-ever, the asteroidal source cannot be confirmed in any of the meteoroids inour sample. Moreover, we also found high dispersion of mineralogical densi-

33

ties among the Taurid meteoroids. Obtained very low densities in individualcases were probably influenced by the high beginning heights of these mete-ors, which relate to the sensitivity of the observational system. Nevertheless,the largest population was identified with mineralogical densities around 2.5g cm-3, which is comparable to the mean values of 2.7 g cm-3 determined byBenyukh (1974) and Babadzhanov & Kokhirova (2009). Mechanical strengthof several Taurids, which exhibited bright flares, was also estimated from thedynamic pressure causing their fragmentation. It was found that studied me-teoroids started fragmenting under low pressures of 0.02 - 0.10 MPa. Whilethese values suggest that Taurid meteoroids are on average weaker comparedto the instrumentally observed carbonaceous meteorite falls (first breakupat 0.1 - 1 MPa), the stream could include bodies of strength comparableto carbonaceous chondrites, as the upper limits of dynamic pressures in oursample reach 0.1 MPa.

Acknowledgements. This work was supported by the Slovak Researchand Development Agency under the contract No. APVV-0516-10, No. APVV-0517-12, and by the Slovak Grant Agency for Science, grants No. VEGA1/0225/14.

References

Asher D. J., 1991, The Taurid meteoroid complex, PhD thesis, New College,Oxford

Asher D. J., Clube S. V. M., Steel D. I., 1993, MNRAS, 264, 93

Asher D. J., Izumi K., 1998, MNRAS, 297, 23

Babadzhanov P. B., 2001, A&A, 373, 329

Babadzhanov P. B., Kokhirova G. I., 2009, A&A, 495, 1, 353

Benyukh V. V., 1968, Vestnik KGU, Seriya astronomii, 10, 51

Benyukh V. V., 1974, PhD. Dissertation Fotometricheskoe issledovanie me-teorov i plotnost meteornych tel, Kiev, 148

Berch et al., 1949, Spravochnik dlya geologov po fizicheskim constantam,Moscow, Inostr. literatura

34

Borovicka J., 1993, A&A, 279, 627

Borovicka J., 1994, Planet. Space Sci., 42, 2, 145

Borovicka J., Spurny P., Keclıkova J., 1995, Astron. Astrophy. Suppl., 112,173

Borovicka J., Koten P., Spurny P., Bocek J., Stork R., 2005, Icarus, 194, 15

Borovicka J., 2006, In Asteroids, Comets, Meteors 2005, ed. D. Lazzaro, S.Ferraz-Mello, J. A. Fernandez, Proc. IAU Symp., 229, 249

Borovicka J., Spurny P., Koten P., 2007, A&A, 473, 661

Borovicka J., Spurny P., Brown P., 2015, Small near-Earth asteroids as asource of meteorites. In Asteroids IV, ed. Michel P., DeMeo F. E., andBottke W. F. Jr., The University of Arizona Press, arXiv:1502.03307

Bronshten V. A., 1983, The Physics of Meteoritic Phenomena, Dordrecht, D.Reidel Publishing Co.

Brown P., Marchenko V., Moser D. E., Weryk R., Cooke W., 2013, Meteorit.Planet. Sci., 48, 270

Brown, P. Wiegert P., Clark D., Tagliaferri E., 2016, Icarus, 266, 96

Ceplecha Z., 1958, Bull. Astron. Inst. of Czech. Ac. of Sci., 9, 4, 154

Ceplecha Z., 1964, Bull. Astron. Inst. of Czech., 15, 102

Ceplecha Z., 1968, SAO Special report, 279

Ceplecha Z., McCrosky R. E., 1976, J. Geophys. Res., 81, 6257

Ceplecha Z., 1987, Bull. Astron. Inst. of Czech., 38, 222

Ceplecha Z., 1988, Bull. Astron. Inst. of Czech., 39, 221

Ceplecha Z., Borovicka J., Elford W., ReVelle D., Hawkes R.L., PorubcanV., Simek M., 1998, Space Sci. Rev., 84, 327

Despois D., 1992, Proc. 150th Symp. Int. Astron. Union, Kluwer, Dordrecht,451

35

Fadeenko Y., 1967, Fizika Gorenija i Vzryva, 3, 276

Haack H., Michelsen R., Stober G., Keuer D., Singer W., Williams I., 2011,Meteorit. Planet. Sci. Suppl., 74, 5271

Haack H., Grau T., Bischoff A., Horstmann M., Wasson J., Sorensen A.,Laubenstein M., Ott U., Palme H., Gellissen M., Greenwood R. C., PearsonV. K., Franchi I. A., Gabelica Z., Schmitt-Kopplin P., 2012, Meteorit.Planet. Sci., 47, 30

Hedin A. E., 1991, J. Geophys. Res., 96, 1159

Hildebrand A. R., McCausland P. J. A., Brown P. G., Longstaffe F. J.,Russell S. D. J., Tagliaferri E., Wacker J. F., Mazur M. J., 2006, Meteorit.Planet. Sci., 41, 3

Hill K. A., Rogers L. A., Hawkes R. L., 2005, A&A, 444, 615

Jenniskens P., Fries M.D., Yin Q.-Z., et al., 2012, Science, 338, 1583

Jones W., Halliday I., 2001, MNRAS, 320, 4, 417

Kikwaya J.-B., Campbell-Brown M., Brown P.G., 2011, A&A, 530, A113, 1

Konovalova N. A., 2003, A&A, 404, 1145

Koten P., Borovicka J., Spurny P., Betlem H., Evans S., 2004, A&A, 428,683

Kresak L., 1978, Bull. Astron. Inst. Czech., 29, 129

Levin B. Y., 1956, Physical Theory of Meteors and Meteor Matter in theSolar System, USSR Academy of Sciences, Moscow

Levison H. F., Terrell D., Wiegert P. A., Dones L., Duncan M. J., 2006,Icarus, 225, 475

Madiedo J. M., Ortiz J. L., Trigo-Rodrıguez J. M., Dergham J., Castro-Tirado A. J., Cabrera-Cano, J., Pujols P., 2014, Icarus, 231, 356

Napier W. M., 2010, MNRAS, 405, 3, 1901

Nagasawa K., 1978, Annals Tokyo Astron. Obs., Second Series, 16, 4

36

Olech A., Zo ladek P., Wisniewski M., Rudawska R., Beben M., KrzyzanowskiT., Myszkiewicz M., Stolarz M., Gawronski M., Gozdalski M., SuchodolskiT., Wegrzyk W., Tyminski Z., 2016, MNRAS, 461, 1, 674

Popescu M., Birlan M., Nedelcu D. A., Vaubaillon J., Cristescu C. P., 2014,A&A, 572, A106

Popova O., Borovicka J., Hartmann W. K., Spurny P., Gnos E., NemtchinovI., Trigo-Rodrıguez J. M., 2011, Meteorit. Planet. Sci., 46, 10, 1525

Porubcan V., Kornos L., Williams I. P., 2006, Contrib. Astron. Obs. SkalnatePleso, 36, 103

Rudawska R., Toth J., Kalmancok D., Zigo P., Matlovic P., 2016, Planet.Space Sci., 123, 25

Shiba Y., 2016, WGN, Journal of the IMO, 44, 3

SonotaCo, 2009, WGN, Journal of the IMO, 37, 55

Southworth R. B., Hawkins G. S., 1963, Smithson. Contrib. Astrophys., 7,261

Spurny P., Borovicka J., Svoren J., Mucke H., 2016, A&A, in preparation

Srirama Rao M., Ramesh P., 1965, Nature, 205, 164

Toth J., Kornos L., Veres P., Silha J., Kalmancok D., Zigo P., Vilagi J., 2011,Publ. Astron. Soc. Japan, 63, 331

Toth J., Kornos L., Zigo P., Gajdos S., Kalmancok D., Vilagi J., Simon J.,Veres P., Silha J., Bucek M., Galad A., Rusnak P., Hrabek P., Duris F.,Rudawska R., 2015, Planet. Space Sci., 118, 102

Trigo-Rodrıguez J. M., Llorca J., 2006, MNRAS, 372, 2, 655

Trigo-Rodrıguez J. M., Llorca J., 2007, Adv. Space Res., 39, 4

Tubiana C., Snodgrass C., Michelsen R., Haack H., Bohnhardt H., Fitzsim-mons A., Williams I. P., 2015, A&A, 584, A97

Vojacek V., Borovicka J., Koten P., Spurny P., Stork R, 2015, A&A, 580,A67, 31

37

Wetherill G. W., 1974, Annu. Rev. Earth Planet. Sci., 2, 303

Wetherill G. W., Revelle D. O., 1981, Icarus, 48, 308

Whipple F., 1940, Proc. Am. Phil. Soc., 83, 711

38